Application of adaptive neuro-fuzzy inference system (ANFIS) for slope and pillar stability assessment

Abstract

This study seeks to apply ANFIS model for stability assessment of surface and underground excavation. Two cases of stability assessment for highway slope and underground rib-pillar mining were performed. This paper presents more than one practical approaches in ANFIS simulation for slope and pillar stability assessment. The most important factors to achieve good performance of ANFIS in this study are types of membership function, number of membership function and number of input variable. The results show that the simulation based on those important factors can reach proper error checking. Considering error checking of the model, type and number of MF simulations can show the best result by comparing more than one simulation. Excellent performance of the 4 (four) input variables in slope case and 3 (three) input variables in pillar case provides valuable information for stability assessment using ANFIS.